Sacubitril/Valsartan in HFpEF: Biomarker Win, Function Gap

Sacubitril/valsartan lowers NT-proBNP, but patients did not walk farther

Sacubitril/valsartan can lower NT-proBNP in heart failure with preserved or mildly reduced ejection fraction. That is a real drug effect, and NT-proBNP is a clinically useful marker of cardiac wall stress. But the best direct randomized evidence here found no parallel improvement in walking distance, symptoms, or quality of life over 24 weeks. A better lab number did not become a better day-to-day result in that trial.

That distinction matters in HFpEF, a syndrome with several overlapping drivers: obesity, hypertension, diabetes, atrial dysfunction, kidney disease, inflammation, and impaired exercise capacity. A falling biomarker can be meaningful without answering the question most patients care about: will I feel better or stay out of the hospital?

NT-proBNP reduction is evidence of a biological response. It is not, by itself, proof that a treatment improves function, quality of life, hospitalization risk, or survival.

What PARALLAX actually found

The central trial is PARALLAX, a randomized clinical trial of 2,572 patients with heart failure and left-ventricular ejection fraction above 40%. It compared sacubitril/valsartan with individualized medical therapy.

At 12 weeks, sacubitril/valsartan reduced NT-proBNP more than the comparator. The adjusted geometric mean ratio was 0.84, with a 95% confidence interval of 0.80 to 0.88. Because that interval stays below 1.0, the biomarker difference was statistically significant.

The patient-centered results were different. At 24 weeks, sacubitril/valsartan did not outperform individualized medical therapy on 6-minute walk distance: the adjusted mean difference was -2.5 meters. It also did not improve Kansas City Cardiomyopathy Questionnaire clinical summary score (KCCQ-CSS), with a mean difference of 0.52 points, or improve New York Heart Association class more than the comparator.

Those neutral findings do not prove that no individual patient can benefit. They do mean that this trial did not show a broad functional or quality-of-life advantage during its 24-week follow-up. The result places a hard limit on claims that lowering NT-proBNP automatically improves how people with HFpEF function.

Safety belongs in the same conversation

In PARALLAX, hypotension occurred in 14.1% of patients assigned to sacubitril/valsartan and 5.5% assigned to comparator therapy. Albuminuria occurred in 12.3% versus 7.6%, and hyperkalemia in 11.6% versus 10.9%. These differences matter in HFpEF, where patients are often older, take several blood-pressure-lowering drugs, and may have limited renal reserve.

Sacubitril/valsartan also has a hemodynamic role. A PARAGON-HF analysis of apparent resistant hypertension found meaningful blood-pressure lowering with sacubitril/valsartan in patients with HFpEF relative to a valsartan-based comparison. That can make the drug useful for selected patients with hypertension, but it also helps explain why blood pressure and tolerability require attention during treatment.

How it compares with other HFpEF drug classes

A 2026 network meta-analysis provides a broader, but indirect, comparison. It pooled 39 trials, 78 treatment arms, and 48,235 patients with HFpEF. Unlike a head-to-head trial, a network meta-analysis estimates relative effects across studies that may differ in populations, comparators, follow-up, and background therapy. Its rankings are useful, but they are not a championship belt.

In that analysis, GLP-1 receptor agonists reduced the composite of cardiovascular death plus heart-failure hospitalization versus placebo, with a hazard ratio of 0.73. SGLT2 inhibitors also reduced that composite, with a hazard ratio of 0.79. GLP-1 receptor agonists ranked first by P-score for the composite and produced the largest estimated gains in 6-minute walk distance and KCCQ-CSS.

Those are morbidity and health-status signals, not a demonstrated survival breakthrough. No therapy in the network meta-analysis significantly reduced cardiovascular death alone or all-cause mortality alone. The practical reading is modest: within this particular network, GLP-1 receptor agonists and SGLT2 inhibitors looked stronger than other classes for the composite outcome, while a mortality-alone advantage remained unproven.

The GLP-1 signal is tied to phenotype

The encouraging incretin evidence should not be generalized to every person labeled HFpEF. A 2025 meta-analysis of randomized trials examined incretin-based therapies specifically in people with obesity and HFmrEF or HFpEF. The STEP-HFpEF program was likewise designed for the obesity-related HFpEF phenotype.

That focus is clinically sensible. Weight loss, insulin resistance, sleep apnea, blood pressure, and exercise limitation often travel together in obesity-related HFpEF. It also means the evidence cannot establish that the same magnitude of benefit applies to lean HFpEF or to every patient with preserved ejection fraction. A real-world study in nonobese patients with type 2 diabetes and HFpEF explored that separate population, but observational comparisons cannot replace randomized evidence for treatment effects.

Why one biomarker cannot summarize HFpEF

HFpEF symptoms arise from more than natriuretic peptide levels. In an observational study of older, obese patients with HFpEF, left atrial stiffness independently predicted exercise intolerance and poorer quality of life. That association does not establish cause and effect, but it illustrates why a treatment can move one marker while leaving walking capacity unchanged.

Inflammation may add another layer. A TOPCAT ancillary study found that higher soluble urokinase plasminogen activator receptor, or suPAR, was associated with worse outcomes in HFpEF. Again, an association is not proof that lowering suPAR would improve outcomes. It does show that the biology is wider than a single natriuretic peptide.

Function can also improve through routes that are not primarily pharmacologic. In a randomized trial that included 59 participants with HFpEF, the HEART Camp exercise intervention improved exercise adherence, physical function, and patient-reported outcomes. The study does not compare exercise with sacubitril/valsartan, but it reinforces a basic point: physical limitation in HFpEF needs direct measurement, not inference from a blood test.

What remains uncertain about sacubitril/valsartan

The available evidence does not support a clean "best HFpEF drug" verdict for sacubitril/valsartan. PARALLAX was strongest for a 12-week biomarker endpoint and did not show functional superiority at 24 weeks. It was not the main PARAGON-HF outcomes report, so it cannot settle effects on heart-failure hospitalization, cardiovascular death, or all-cause mortality across narrower ejection-fraction bands.

Population differences also complicate broad conclusions. A PARAGON-HF analysis of global differences reported substantial regional variation in patient characteristics and comorbidities. Such differences can affect event rates, baseline risk, treatment tolerance, and apparent treatment effects.

Real-world data can add context, but not certainty. A 2024 cohort study comparing sacubitril/valsartan with angiotensin receptor blockers examined hospitalization outcomes in routine practice. Because treatment was not randomized, differences between groups may remain even after statistical adjustment.

The practical verdict

Sacubitril/valsartan has credible randomized evidence for lowering NT-proBNP in heart failure with ejection fraction above 40%. In PARALLAX, it did not show a matching benefit in 6-minute walk distance, KCCQ-CSS, or NYHA class over 24 weeks, and it increased hypotension.

For comparative HFpEF treatment signals, the 2026 network meta-analysis placed GLP-1 receptor agonists and SGLT2 inhibitors ahead on the composite of cardiovascular death plus heart-failure hospitalization. The GLP-1 evidence is most relevant to obesity-enriched HFmrEF/HFpEF populations, and neither class showed a statistically significant reduction in cardiovascular death alone or all-cause mortality in that analysis.

Patients and clinicians should judge treatment by the outcomes that matter for the individual: symptoms, exercise tolerance, blood pressure, kidney function, hospitalization risk, adverse effects, and the phenotype driving the HFpEF. NT-proBNP can inform that discussion. It cannot finish it.